[Laser] Prism Tuner

[Tim Toast]

Hi All,

Seeing some talk about narrow optical filters recently, i wanted to bring up an old idea 
about using a glass prism to tune or pre-select the optical passband.

here's a diagram: http://www.aladal.net/toast/prismtuner1.jpg

One possible advantage of a prism over a color filter may be that the prism is unaffected by
the large angles of entrance of the beam from a large lens. Also an advantage of a prism over
a transmission grating (or reflective grating) is that the prism produces only one "order" or
spectrum, and all the received power is in one spectrum. A grating would divide up the power 
between several orders.

So, we have a converging beam from the lens passing through the prism and on to the focal
plane of the photodiode. The lens system still produces an image at the focal plane but it is
spread out in one dimension according to frequency. So, when viewing a white point source, the
receiver would produce a narrow line in the focal plane with blue at one end and red at the 
other. 
-R----G-----B-
Placing the prism just in front of the focal point, in effect, turns the receiver into a full 
aperture or slitless spectrograph. Instead of a "normal" spectrograph where you have a very 
narrow entrance slit that the main lens is focused on, you take away the slit and focus the 
lens directly on the detector via the prism.

With just a short distance between the prism and the focal points, there isn't much room for the
spectrum to spread out and the focal line will be relatively short. There might only be a few
millimeters between the red and blue focal points. And this would be especially true with short
focal length systems common in light beam coms.

One interesting thing about this is that the bandwidth is mostly determined by the physical 
diameter of the sensitive area of the photodiode used. Ignoring the nonlinear nature of the 
prism's spectrum for a moment - the distance between the red and blue focus divided by the 
diameter of the photodiode equals the bandwidth. So if you have a diode with a diameter of 
one millimeter and the focal line from the prism is 10mm in length, then the bandwidth is 
about 1/10th of the full 300nm wide spectrum or 30nm.  In practice, the spectrum produced 
from a prism is compressed toward the red end and expanded toward the blue end. So with a 
fixed size photodiode, the bandwidth would be narrower at the blue end and wider at the red 
end. But you could ignore that for the most part and just calibrate the tuning scale by 
using known wavelength light sources. 

So, i hope this comes in handy somehow. It may be that there are some factors that make 
regular color filters better than a prism in the long run. The one that worries me the most 
is the short focal lengths used commonly because there's a limit for how steep the angles 
can be and still not block any light from the edges of the beam by the prism, while having the 
focal points available outside of it. 
All this assumes there's a need for cheap, quick or continuous tunability. If nothing else, 
it could be useful as a wavelength measuring device that can also be used in the IR or UV 
with a broadbanded photodiode.

- toast